Professional Documents
Culture Documents
Download pdf Advances In Optoelectronic Technology And Industry Development Proceedings Of The 12Th International Symposium On Photonics And Optoelectronics Sopo 2019 August 17 19 2019 Xi An China 1St Edition Gin ebook full chapter
Download pdf Advances In Optoelectronic Technology And Industry Development Proceedings Of The 12Th International Symposium On Photonics And Optoelectronics Sopo 2019 August 17 19 2019 Xi An China 1St Edition Gin ebook full chapter
https://textbookfull.com/product/frontiers-in-cyber-security-
second-international-conference-fcs-2019-xi-an-china-
november-15-17-2019-proceedings-bazhong-shen/
https://textbookfull.com/product/machine-learning-for-cyber-
security-second-international-conference-ml4cs-2019-xi-an-china-
september-19-21-2019-proceedings-xiaofeng-chen/
https://textbookfull.com/product/proceedings-of-the-12th-
international-symposium-on-computer-science-in-sport-
iacss-2019-martin-lames/
https://textbookfull.com/product/report-on-the-development-of-
cruise-industry-in-china-2019-hong-wang/
Recent Advances in Information and Communication
Technology 2019: Proceedings of the 15th International
Conference on Computing and Information Technology
(IC2IT 2019) Pongsarun Boonyopakorn
https://textbookfull.com/product/recent-advances-in-information-
and-communication-technology-2019-proceedings-of-the-15th-
international-conference-on-computing-and-information-technology-
ic2it-2019-pongsarun-boonyopakorn/
https://textbookfull.com/product/dependable-software-engineering-
theories-tools-and-applications-5th-international-symposium-
setta-2019-shanghai-china-november-27-29-2019-proceedings-nan-
guan/
https://textbookfull.com/product/advanced-parallel-processing-
technologies-13th-international-symposium-appt-2019-tianjin-
china-august-15-16-2019-proceedings-pen-chung-yew/
https://textbookfull.com/product/e-learning-e-education-and-
online-training-5th-eai-international-conference-
eleot-2019-kunming-china-august-18-19-2019-proceedings-guanglu-
ADVANCES IN OPTOELECTRONIC TECHNOLOGY AND INDUSTRY
DEVELOPMENT
Symposium on Photonics and Optoelectronics:
proceedings
The International Symposia on Photonics and Optoelectronics (SOPO) aim
to provide a premier technical forum for researchers, engineers as well as
professionals from all over the world to present the latest research and devel-
opment in Photonics and Optoelectronics related fields. Optoelectronic sys-
tems use laser technology, for a wide range of applications in the fields of
telecommunications, electronic industrial, civil, and aeronautical engineering.
SOPO features the latest advances in optoelectronic integration, laser tech-
nology, manufacturing of optical components, advanced laser materials,
films and devices, optical communications, silicon photonics, quantum optics,
optoelectronic devices and integration, medical and biological applications
and image processing.
Advances in Optoelectronic
Technology and Industry
Development
Edited by
Gin Jose
University of Leeds, UK
Mário Ferreira
University of Aveiro, Aveiro, Portugal
CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business
All rights reserved. No part of this publication or the information contained herein may be
reproduced, stored in a retrieval system, or transmitted in any form or by any means,
electronic, mechanical, by photocopying, recording or otherwise, without written prior
permission from the publisher.
Although all care is taken to ensure integrity and the quality of this publication and the
information herein, no responsibility is assumed by the publishers nor the author for any
damage to the property or persons as a result of operation or use of this publication and/or
the information contained herein.
Table of contents
Preface ix
Organizers xi
Image Processing
Electrically-modulated optoelectronics-based infrared source enabling ground surface
precision deflectometry 3
H. Quach, L.R. Graves, H. Kang & D.W.Kim
Adaptive learning rate and target re-detection for object tracking based on correlation filter 13
J. Xiang & P. Shen
Calculation method of infrared temperature on the natural ground surface 20
C. Shan & M. Jun-chun
Research on gesture-recognition method in video based on the sparse representation theory 25
Y. Lei, L. Feng, L. Zhenglong & W. Shiliang
v
Optical Communications
Performance investigation of 16/32-channel DWDM PON and long-reach PON systems
using an ASE noise source 93
D.S. Sundar, T. Sridarshini, R. Sitharthan, M. Karthikeyan, A.S. Raja & M.F. Carrasco
A comparative selection of the low-loss optical fibers designed for FTTH networks 100
F.E. Seraji, A. Emami & D.R. Rafi
Enhancement of fidelity of quantum teleportation in a non-Markovian environment 109
Y. Zhangy & X. Wu
A 2×2 optical switch based on semiconductor optical amplifier cross-gain modulation
technology 117
S. Zhou, Z. Qi, D. Ding & D. Deng
A PSK quantum-noise randomized cipher simulation system model based on standard
commercial devices 125
Y. Chen, T. Pu, Hua Zhou, H. Shi, H. Tang, Y. Li, H. Jiao & H. Zhang
Propagation characteristics of Super-Gaussian pulse in dispersion-decreasing fiber 131
S. Shi & Q. Zhang
Time-delay measurement of optical-fiber link based on time-frequency simultaneous
transmission method 140
J.C. Guo, L. Lu, H. Wei & X.Y. Zhao
SNR uniformity optimization for LEDs ring alignment in visible light communications 146
F. Li, J. Xiao, X. Li, J. Xia, Q. Zhou & G. Hu
vi
A low-cost and compact fiber-optic sensor based on modal interference for humidity sensing 197
Y. Liu, P. Li, N. Zhang & X. Li
vii
Advances in Optoelectronic Technology and Industry Development – Jose & Ferreira (eds)
© 2020 Taylor & Francis Group, London, ISBN 978-0-367-24634-1
Preface
The fields of Photonics and Optoelectronics have grown spectacularly over the last few dec-
ades. Actually, these fields are intertwined and have a profound effect on the emergence of
modern technologies and their influence on our lives. Experimental and theoretical research
into such technologies include sources and detectors as components, active elements such as
modulators/switches, integration of optical components to perform various functions, and the
development of active optical system/subsystem technologies.
SOPO 2019 featured the latest advances in optoelectronic integration, laser technology, manu-
facturing of optical components, advanced laser materials, films and devices, optical commu-
nications, silicon photonics, quantum optics, optoelectronic devices and integration, medical
and biological applications and image processing. The conference included 22 Plenary
Speeches and many other regular communications. However, due to Ei indexing requirements,
only a limited number of papers have been selected for publication. Actually, the 36 papers
included in this Proceedings provide an excellent overview of the topics presented at SOPO
2019. Among these papers, 8 are dedicated to optical sensors, 7 are in the area of optical com-
munications, 7 are related with imaging and detection, 2 are in the area of quantum optics, 6
are concerned with lasers and other optical sources, while the remaining 6 papers are dedi-
cated to different types of optoelectronic and photonic devices. We are sure that this publica-
tion will arouse great interest both from specialists in these fields and from general readers.
SOPO 2019 could not have happened without the hard work of many people, namely those
integrating the Local Organizing Committee and the Session Chairs, who helped in many
ways to assemble and run the conference, the Technical Program Committee and, especially,
the General Chair, Prof. Zhiping Zhou, who guided the technical direction of the conference
and assisted with the conference program and proceedings. We are also extremely pleased to
have received generous sponsorship from several organizations. Finally, we are also very
grateful to the Editorial Staff at CRC Press/Balkema, and the many contributors and
reviewers for helping us to put-together this Proceedings
Mário F. S. Ferreira
Main Editor
Department of Physics, University of Aveiro, Aveiro, Portugal
ix
Advances in Optoelectronic Technology and Industry Development – Jose & Ferreira (eds)
© 2020 Taylor & Francis Group, London, ISBN 978-0-367-24634-1
Organizers
Wuhan University
Beijing University of Posts and Telecommunications
IPOC State Key Lab
Liaocheng University
GENERAL CHAIR
xi
Islamic University of Gaza, Palestine
• Prof. Zheng Shang Da
Xi’an Institute of Optics & Precision Mechanics of Chinese Academy of Sciences, China
• Dr. Kang Wei
Apple, Inc, USA
• Prof. Gin Jose
University of Leeds, UK
• Dr. Te Hu
Apple Camera Team, USA
• Dr. Lin Xu
University of Southampton, UK
• Prof. Jean-Luc ADAM
UMR CNRS 6226 Institut des Sciences Chimiques de Rennes, France
• Dr. Xiaotian Li
Grating Technology Laboratory, Changchun Grating Technology Laboratory, Chinese
Academy of Sciences, China
• Dr. Gholamreza Shayeganrad
University of Southampton, UK
• Prof. Khatereh Khorsandi
Tehran University of medical sciences (TUMS) branch, Iran
xii
Image processing
Advances in Optoelectronic Technology and Industry Development – Jose & Ferreira (eds)
© 2020 Taylor & Francis Group, London, ISBN 978-0-367-24634-1
Henry Quach
James C. Wyant College of Optical Sciences, University of Arizona, Tucson, AZ, USA
Logan R. Graves
James C. Wyant College of Optical Sciences, University of Arizona, Tucson, AZ, USA
Intuitive Optical Design Lab, Tucson, AZ, USA
Hyukmo Kang
James C. Wyant College of Optical Sciences, University of Arizona, Tucson, AZ, USA
1 INTRODUCTION
3
custom null optics are required. Diffractive elements called computer-generated holograms
(CGH) are attractive null element options because they can null up to and including arbitrary
surfaces with extreme or high-frequency features (Dubin et al., 2009). However, a CGH is typic-
ally designed for only one null configuration and may not always be a viable option for testing.
Alternatively, deflectometry is a non-null test method that has demonstrated comparable
performance to interferometry over a range of freeform surfaces (Graves et al., 2019). In this
method, an illumination source emits a known pattern, which specularly reflects off the UUT,
and the reflected image of the source is captured by a camera. With precise knowledge of the
setup’s geometry, the local slopes of the UUT can be determined and integrated to generate
a reconstructed surface map. Deflectometry carries the advantages of high testable slope
dynamic range and measurement flexibility because it does not require a null reference. Of
interest to optical fabrication, measurement of surface figure error during intermediary grind-
ing phases may help prioritize local figure error correction so that a surface may more effi-
ciently converge. Due to rapidly-changing surface profiles during grinding, CGHs are not
practical, but deflectometry can become applicable if a camera and source of appropriate
wavelength satisfy the specular reflectance condition (Oh et al., 2016; Lowman et al., 2018).
For optics in the grinding phase, which have a rough diffusing surface, a specular reflection
can be achieved using long-wave infrared (LWIR) light, specifically, in the 7–14 µm region.
LWIR cameras are readily obtainable, but LWIR thermal source options are currently limited.
One common design is implemented by applying a current to a thin tungsten ribbon, which
induces joule heating and creates a rectangular, pseudo-blackbody emitting source (Su et al.,
2011, 2013; Oh et al., 2016). By scanning this rectangular ribbon in orthogonal directions, a line
scanning source is created, and slope information can be obtained to construct a full aperture
surface sag map. While this source enables infrared deflectometry testing of rough surfaces at
a precision scale, several inherent characteristics including low signal-to-noise ratio (SNR),
limited modulation depth, and low temporal stability, limit the testing accuracy and range.
We have created a new source design which addresses these prior issues and extends the
applicable range of infrared deflectometry testing. The source is a Long-wave Infrared Time-
Modulated Integrating cavity Source (LITMIS), which uses modular high efficiency and high
stability resistive membrane pseudo-blackbody elements. A performance comparison was
made against the traditional tungsten ribbon source, whose shape was identical to the exit slit
of the box, using the same setup and LWIR camera. Results show promise in deflectometric
testing of rough optics during their grinding phase and UUTs under thermal load.
4
extended to all camera pixels to measure the local slopes at all mirror pixels on the UUT in
orthogonal directions, referred to as SX(x,y,z) and SY(x,y,z), representing the x and y slopes
respectively. These slope maps are typically integrated with a zonal integration method such as
Southwell integration (Southwell, 1980) or a modal integration such as using a gradient Cheby-
shev polynomial set (Aftab et al., 2018), resulting in a reconstructed surface map. Figure 1 dem-
onstrates a standard deflectometry setup and the model used for local slope calculation.
Figure 1. In a standard deflectometry system (a), light is emitted by a source, s(x,y,z). Some of the
emitted light will deflect from the UUT surface, u(x,y,z), and be successfully captured by the
camera, c(x,y,z), with an entrance pupil location, p(x,y,z). Represented by the vector ~ v(x,y,z),
a deflected light ray follows the law of reflection and is reflected by a local area on the UUT surface,
whose normal vector is given by ~ n. Knowing the ray start, intercept, and end location, the local slope
of the UUT, given by SX, can be determined (b).
5
of a thin ribbon to possibly experience any of its lower bending modes during testing, further
deviating the emission area from the theoretical ideal rectangular area. Additionally, driving
higher input current through the ribbon only accelerates material degradation and also shifts
the output spectrum away from desirable longer wavelengths. Finally, the heated tungsten
ribbon invariably creates a temperature gradient in the surrounding air due to local convective
boundary conditions. For larger ribbons and longer source operation time, growing pockets
of locally heated air may generate thermal noise that blurs the infrared source signal.
We propose an alternative source called a Long-wave Infrared Time-modulated Integrating
cavity Source (LITMIS) to address these limitations. At the core of the LITMIS design is
modular usage of EMIRS200 Series emitters, which are available commercially off-the-shelf
from Axetris AG. The packaged emitters, henceforth referred to as ‘caps’, are typically used
in nondispersive infrared (NDIR) and photoacoustic gas spectroscopy and have a usable life-
time of 10,000 hours (Esfahani & Covington, 2017; Wilson et al., 2019). Although the emissive
spectral distribution is centered at 4 µm at the operational temperature (550°C), radiant flux
across the desired 7-12 µm band of interest is also produced, as described by Planck’s Law.
Caps are fabricated by electroplating Platinum Black, a platinum powder, onto a thin mem-
brane which floats on a silicon substrate (Hessler et al., 2004; Axetris AG, 2014, 2019). Similar
to the tungsten ribbon, a cap is operated by driving an electric current through the target
material to induce resistive heating and emit blackbody radiation. Two desirable source prop-
erties result from the cap material design and geometry:
• High emissivity: when electroplated onto a substrate, Platinum Black forms dendritic
structures on the scale of < 50 nm. The feature scale and high porosity of the structure
result in high absorptivity and high emissivity (ε > 0.9) characteristics for the coated mem-
brane, unlike that of smooth metallic surfaces. For example, the typical emissivity of tung-
sten is ε < 0.1 at 1000K (Verret & Ramanathan, 1978).
• Low thermal capacitance: this property, also called thermal mass, refers to a body’s ability
to store thermal energy. Here, the combined resistive network of an exterior Platinum
Black coating, Tantalum Oxide, and a SiN membrane is only several microns thick (Hessler
et al., 2004), as compared to a typical ~30 µm thick tungsten ribbon. Because the ultra-low
mass is suspended in air by a silicon substrate, the thin emissive network can heat to and
cool from its steady-state temperature in less than 1 second (Hessler et al., 2004).
Temporal modulation emerges as a viable operational mode of the new source. Testing
reveals that a cap can achieve an 80% contrast ratio at 1 Hz, which enables in-situ background
noise images to be taken during testing. Therefore, at a given image capture, a pair of ‘signal’
and ‘background’ shots can be successively captured for noise subtraction in post-processing.
Figure 2. The Axetris EMIRS200 TO39 source features a resistive membrane that, when current is
applied, acts as a pseudo-blackbody source (a). The emitters can be connected in series or in parallel,
much likea group of resistors, to create a patterned source (b).
6
This option is not practical for the tungsten ribbon source because its relatively higher thermal
mass evinces a slower thermal transient response. For any thermal source, noise will accumulate
in the environment as it warms up and approaches its steady-state operational temperature.
Figure 3. The final optimized LITMIS assembly was modeled in SolidWorks (a). 20 infrared emitters
point inside a diffusing cavity which contains a single exit slit, machined from a thin aluminum plate. The
optimized design was also modeled in LightTools, where the irradiance at the surface of the box was
simulated to assure high uniformity across the exit slit (b).
7
a 160 × 120 microbolometer array with 320 × 240 super-resolution output. Exposure, gain,
and level settings were adjusted to prevent output saturation and were held constant between
all tests. A fixed optical mount was situated approximately one meter from the camera and
allowed for repeatable UUT placement. To compare properties between source modalities,
a scanning platform was utilized with a mounting interface to interchange LWIR sources.
A motorized lead screw stage moved source assemblies in the vertical direction with an abso-
lute positional accuracy of ±0.005 mm. For ease of comparison, sources shared both identical
slit dimensions (75 mm × 2.5 mm) and radiant exitance planes. Figure 4 demonstrates the
camera and source setup for the test system.
UUTs included a 2-inch diameter rough ground glass referred to as Glass1500 and
a bare aluminum flat referred to as AlRoom. UUTs were selected for measurement due
to their diffuse nature, making thermal infrared deflectometry an ideal metrology
method. When heated to 150°C, AlRoom flat is referred to as Al150, and is done so as
a challenge case because it generates variable thermal background noise, which typically
degrades the SNR of a test. This is a common scenario, as several optics, such as solar
collectors or even the DKIST primary mirror, will operate under thermal load; thus,
the Al150 test case is highly relevant. The surface roughness of both optics was meas-
ured using a Zygo NewView 8300 Interference Microscope. The ground glass surface
featured a surface roughness of 127.89 nm RMS while the bare aluminum surface
roughness was 102.53 nm RMS over a small 834 × 834 µm square area over each optic.
The LWIR line source was implemented by running direct current (2.2 A, 2.1 W)
across a thin 32 μm rectangular tungsten ribbon. Transient thermal noise from the wire,
such as the local heating of air, was reduced by taking measurements only after the
ribbon reached a thermal steady-state condition, which was achieved after approximately
20 minutes.
The LITMIS source was implemented by applying a (0.29 A, 7 W) load to a circuit
consisting of 20 emitters. Pointed into the enclosure, the rectangular emitter array was
operated by binary power cycling with a digitally controlled relay. Enclosure walls were
machined from bare Al 6061-T6 and characterized to 3.4 μm RMS by a Zygo NewView
8300 Interference Microscope. Lastly, to minimize latent thermal radiation from the
cavity interior, the LITMIS source was cooled to 0°C prior to each test. Figure 5 dis-
plays the integrating cavity after assembly.
3.2 Measurements and results for source geometry and temporal stability
Geometrical profile measurements were conducted by recording an image focused at each
source with the LWIR camera. The camera software was adjusted so the maximum 8-bit signal
count was a discrete value just below 255. Source profiles were captured and compared against
Figure 4. A traditional tungsten ribbon scanning source is mapped to Thermal-Eye 3500AS camera. As
a requirement for deflectometry measurement, the camera is focused on the surface of the UUT. LWIR
radiation leaves the source, of which, some rays are intercepted by the detector. For ground optics,
higher SNR is required to precisely, or at all, determine the surface slope at a given mirror pixel. After
UUT measurement with the ribbon, it is unmounted and replaced by the LITMIS source.
8
Figure 5. A frontal view of the assembled LITMIS architecture is shown in (a). Four parallel
chains of elements are connected to operate as a single electrical circuit. While radiative thermal
input is intended towards the interior of the box, excess thermal radiation is emitted by the heated
copper body of each cap. Aluminum-covered shielding is mounted to the LITMIS façade to deflect
it, as shown in (b). By design, LWIR radiation does not propagate directly from a heat source to
the object of measurement. Rather, it is diffused by internal scattering and may only escape
through a limited window. With this mechanism, input power can be readily scaled by simply
adding more emitters within the cavity. In contrast, the available input power from a direct ther-
mal source is limited by the area of its profile geometry.
to an assumed ideal rectangular emission profile. As seen in Figure 6, the LITMIS source main-
tains a flatter signal power across the horizontal pixel band and drops off sharply at the edges.
To observe the temporal stability, a measurement was performed by focusing the
camera on each source, which was turned on and recorded for 30 minutes, separately.
Temporal measurement results illustrate the high stability and uniform slit emission
Figure 6. The images of the tungsten ribbon (a) as well as the LITMIS slit (b) sources were cap-
tured using Thermal-Eye 3500AS camera through-focused onto the source through a flat mirror.
Observing the camera signal power histograms for the tungsten ribbon (c) and the LITMIS slit
(d) sources, it is seen that the average power is similar, but the source profile geometries are
quite different, where both should ideally form a flat top rectangular shape with roll-on and
roll-off at the edges. This is due to the convolution of the rectangular source with the circular
camera pupil.
9
Figure 7. Using both a traditional tungsten ribbon source (top row) and the LITMIS source
(bottom row), infrared deflectometry measurements were taken and the surface reconstructed for
the Glass1500 optic (left column), the AlRoom optic (middle column), and the Al150 optic (right
column). For all maps, Standard Zernike terms 1:37 were removed to observe the surface mid-to-
high spatial frequency topology. In the Glass1500 case, some points in the aperture could not be
sampled due to insufficient SNR. In the Al150 case, a majority of the surface topology was unavail-
able due to competitive thermal noise. With greater immunity from thermal noise, the LITMIS
source obtained sufficient slope data to reconstruct all three samples.
geometry of the LITMIS source. Stability was most pronounced in the signal peak-to-
valley fluctuation comparison, which was measured as 11 signal counts for the tungsten
ribbon source and 1.82 for the LITMIS source.
3.3 Measurements and results for UUT surface reconstruction and repeatability
To determine the comparative surface reconstruction repeatability, the Glass1500,
AlRoom, and Al150 optics were measured using both tungsten ribbon and LITMIS
sources.
The reconstructed surface topology maps of all 3 UUT cases are shown in Figure 7.
Both the LITMIS slit source and tungsten ribbon source were successfully used to test
the Glass1500 and AlRoom optics. However, the standard deviation of the signal power
across the five repeat measurements performed for every optic for each source was
slightly larger for the tungsten ribbon as compared to the LITMIS source. Despite this,
the LITMIS source was better able to reduce noise for all cases, which directly impacts
the SNR of both test methods. Overall, the LITMIS source achieved a 2–5 times larger
SNR for the Glass1500 and AlRoom. The Al150 sample could not be measured using the
ribbon source due to high thermal noise generated by the UUT, while the LITMIS
source provided sufficient slope data for surface reconstruction.
4 CONCLUSIONS
In infrared deflectometry, any uncertainty in the spatial and temporal behavior of a source
directly negatively impacts the reconstruction accuracy and uncertainty. Additionally, time-
varying, thermal background noise is common in most test environments and degrades the
effectiveness of LWIR sources by decreasing the signal-to-noise ratio. We have instead created
an integrating cavity source, which emits long-wave infrared light uniformly from a defined
exit slit, which we call the LITMIS source. A demonstration infrared deflectometry system
10
using the new source successfully tested a ground glass sample and aluminum blank, as well as
a previously unmeasurable aluminum blank under thermal load. In all cases, the source exhib-
ited excellent repeatability and significantly improved the SNR of the test, as compared to
testing using a traditional tungsten ribbon. The authors encourage readers to examine further
detailed test results and analysis in the paper, “High-contrast thermal deflectometry using long-
wave infrared time-modulated integrating cavity.” (Graves et al., in prep.)
The scalability and flexibility of the new source architecture leave much to explore. Beyond
simply engineering the source design to achieve high signal output power, temporal behavior
may be exploited to reconstruct shapes of challenging optics. As temporal modulation adequately
inoculates measurement from local thermal fluctuations in active workpieces, the creation of an
on-machine deflectometry system also edges towards the realm of possibility. More excitingly,
these building blocks may be applied towards a sinusoidally modulated infrared source, which
may allow phase-shifting infrared deflectometry to become viable in optical metrology.
ACKNOWLEDGMENTS
The authors would like to acknowledge the II-VI Foundation Block-Gift Program for helping
support general deflectometry research in the LOFT group, making this research possible.
Also, this work was made possible in part by the Technology Research Initiative Fund Optics/
Imaging Program and the Korea Basic Science Institute Foundation.
REFERENCES
Aftab, M., Burge, J.H., Smith, G.A., Graves, L.R., Oh, C.J. and Kim, D.W. 2018. ‘Chebyshev gradient
polynomials for high resolution surface and wavefront reconstruction’, 1074211(September 2018),
p. 40. doi:10.1117/12.2320804.
Axetris AG 2014. ‘IRS Frequently Asked Questions Rev.A’, pp. 2–5.
Axetris AG 2019. ‘Infrared Sources Brochure (English)’, pp. 42–75. doi:10.1201/9780203750834-3
Dubin, M.B., Su, P. and Burge, J.H. 2009. ‘Fizeau interferometer with spherical reference and
CGH correction for measuring large convex aspheres’, Optical Manufacturing and Testing VIII, 7426
(May), p. 74260S. doi:10.1117/12.829053.
Esfahani, S. and Covington, J.A. 2017. ‘Low Cost Optical Electronic Nose for Biomedical Applications’,
Proceedings, 1(10), p. 589. doi:10.3390/proceedings1040589
Fang, F.Z., Zhang, X.D., Weckenmann, A., Zhang, G.X. and Evans, C. 2013. ‘Manufacturing and meas-
urement of freeform optics’, CIRP Annals - Manufacturing Technology. CIRP, 62(2), pp. 823–846.
doi:10.1016/j.cirp.2013.05.003.
Graves, L.R., Quach, H., Koshel, J.R., Oh, C.J., Kim, D.W. (in prep). ‘High contrast thermal deflecto-
metry using long-wave infrared time-modulated integrating cavity.’
Graves, L.R., Quach, H., Choi, H. and Kim, D.W. 2019. ‘Infinite deflectometry enabling 2π-steradian
measurement range’, Optics Express, 27(5), p. 7602. doi:10.1364/oe.27.007602.
Hessler, T., Dubochet, O., Forster, M. and Merschdorf, M. 2004. ‘Micro-machined, electrically modu-
lated thermal infrared source with black body characteristic pyroelectric detectors or’,
(November 2015).
Höfer, S., Burke, J. and Heizmann, M. 2016. ‘Infrared deflectometry for the inspection of diffusely specu-
lar surfaces’, Advanced Optical Technologies, 5(5–6), pp. 377–387. doi:10.1515/aot-2016-0051.
Kim, D.W., Su, T., Su, P., Oh, C., Graves, L. and Burge, J. 2015. ‘Accurate and rapid IR metrology for
the manufacture of freeform optics’, SPIE Newsroom, pp. 9–11. doi:10.1117/2.1201506.006015.
Kim, D.W., Oh, C., Lowman, A., Smith, G.A., Aftab, M. and Burge, J.H. 2016. ‘Manufacturing of
super-polished large aspheric/freeform optics’, Advances in Optical and Mechanical Technologies for
Telescopes and Instrumentation II, 9912, p. 99120F. doi:10.1117/12.2232237.
Lowman, A.E., Yoo, H., Smith, G.A., Oh, C.J. and Dubin, M. 2018. ‘Improvements in the scanning
long-wave optical test system’, 1074216(September 2018), p. 48. doi:10.1117/12.2321265.
Martin, H.M. et al. 2018. ‘Manufacture of primary mirror segments for the Giant Magellan Telescope’,
(May), p. 30. doi:10.1117/12.2312935.
Oh, C.J. et al. 2016. ‘Fabrication and testing of 4.2m off-axis aspheric primary mirror of Daniel K. Inouye
Solar Telescope’, 9912, p. 99120O. doi:10.1117/12.2229324.
11
Southwell, W.H. 1980. ‘Wave-front estimation from wave-front slope measurements’, Journal of the
Optical Society of America, 70(8), p. 998. doi:10.1364/josa.70.000998.
Su, T., Park, W.H., Parks, R.E., Su, P. and Burge, J.H. 2011. ‘Scanning Long-wave Optical Test System:
a new ground optical surface slope test system’, Optical Manufacturing and Testing IX, 8126(May),
p. 81260E. doi:10.1117/12.892666.
Su, T., Wang, S., Parks, R.E., Su, P. and Burge, J.H. 2013. ‘Measuring rough optical surfaces using scan-
ning long-wave optical test system 1 Principle and implementation’, Applied Optics, 52(29), p. 7117.
doi:10.1364/ao.52.007117.
Verret, D.P. and Ramanathan, K.G. 1978. ‘Total hemispherical emissivity of tungsten’, Journal of the
Optical Society of America, 68(9), p. 1167. doi:10.1364/JOSA.68.001167.
Wilson, D., Phair, J. and Lengden, M. 2019. ‘Performance Analysis of a Novel Pyroelectric Device for
Non-Dispersive Infra-Red CO2 Detection’, IEEE Sensors Journal, 1748(c), pp. 1–1. doi:10.1109/
jsen.2019.2911737.
12
Advances in Optoelectronic Technology and Industry Development – Jose & Ferreira (eds)
© 2020 Taylor & Francis Group, London, ISBN 978-0-367-24634-1
ABSTRACT: In this paper, two problems about the update rate and long-term tracking in
target tracking model are discussed. Traditional correlation filter tracker only uses a fixed rate
mechanism, so the target update rate is fixed. In this paper, we improved it so that it can adjust
the update rate adaptively according to the similarity between different image sequences and first
frame. Besides, in order to deal with more challenging scenarios and long-term tracking targets,
we add a re-detection mechanism to the tracker. This method overcomes the limitation of the
traditional correlation filter tracker using fixed update rate by studying the similarity between the
frames of the image, and can adaptively change the update rate of the model. A large number of
experimental results show the superiority of our improved tracker in accuracy and success rate.
1 INTRODUCTION
Object tracking is the core of computer vision and has been widely used in surveillance,
human-computer interaction, robotics and other fields Wibowo,2018).In recent years, the
tracker based on kernel correlation filter has excellent performance and it can maintain high
frame rate with high accuracy, and the current mainstream benchmark rankings are very high.
However, most of the kernel correlation filtering algorithms adopt a fixed learning rate, which
cannot adapt to the change of tracking environment. Moreover, most short-term tracking has
no target re-detection mechanism, and cannot deal with the long-term tracking problem well.
The proposed method is closely related to the correlation filter-based tracker which applies
the correlation filter in traditional signal processing technology to tracking applications.
Visual tracking has been studied extensively with numerous applications (Smeulders, 2014). In
2010, Bolme et al. (2010) proposed a Minimum Output Sum of Squared Error (MOSSE)
filter. It is the first time to apply the correlation filter to the tracking field. The MOSSE is
simple in calculation and can track the target quickly, but it cannot guarantee accurate
tracking when the appearance of the target changes. In 2014, Li et al. fused the features of
color-naming (CN) (van de Weijer, 2009) and histogram of orientation gradients (HOG) and
predicted the scale change of target by using scale pool. Then, the Scale Adaptive with Mul-
tiple Features Tracker (SAMF) was proposed (Li, 2014). SAMF improved the discriminant
ability of target on the basis of High-speed tracking with kernelized correlation filter (Henri-
ques, 2015) and solved the scale change problem to a certain extent. It has good tracking
effect, but it still needs to be improved for fast moving target and partly occluded target track-
ing. In 2015, Ma Chao et al. proposed LCT algorithm which takes into account the context
and scale transformation of the target (Ma, 2015). Based on the correlation filtering,
a random fern-based re-detector is combined to further improve its long-term tracking effect.
The traditional tracking framework has boundary effect. Mueller et al. (2017) proposed
a Context-Aware (CA) framework to solve the boundary effect caused by cyclic sampling
which can be integrated with many classical CF trackers. Tracking-by-detection trackers have
attracted wide attention due to their high performance and efficiency (Kalal, 2012).
In this paper, we mainly study single target tracking problems. We consider the problems men-
tioned above and propose a novel Adaptive learning rate and re-detection tracker (ALRD). The
13
proposed method overcomes the limitation of the traditional correlation filter tracker using fixed
update rate by studying the similarity between each frame of the image and the first frame to adap-
tively change the update rate of the model. Meanwhile, in order to deal with more challenging
scenarios and long-term tracking targets, we have added a re-detection mechanism to the tracker.
The main contributions of this paper are summarized below: (1) The fixed update rate of
the tracker based on correlation filter is improved so that it can adapt the update rate on the
basis of the different image sequences; (2) A re-detection mechanism is added to the tracker so
that the tracker can track the target for a long time; (3) A large number of experiments have
been carried out to compare the improved tracking algorithm and some short-term tracking
algorithm with our proposed method, which includes adaptive update rate and increasing re-
detection mechanism to cope with long-term tracking, and the results show the better per-
formance of the proposed tracker in terms of accuracy and success rate.
2 RELATED WORK
The tracker proposed in this paper is based on SAMF tracking of CA framework. In the fol-
lowing passage, we use SAMF(CA) to represent based on SAMF tracking of CA framework.
We introduce the differential hashing algorithm in order to solve the update rate is a fixed
value in target tracking this problem.
X ¼ FH diagðFxÞF ð1Þ
where F is the DFT matrix and FH is the Hermitian transpose of F. The cyclic structure of this
matrix helps to effectively solve the problem of regression in the Fourier domain ridge.
Here, the learning correlation filter is represented by a vector w. A0 represents the cyclic
shift matrix, and the regression target is y.λ1 is a regularization factor parameter.
Because the surrounding environment of the tracked object has a great impact on the tracking
performance of the tracker, for example, there is a lot of background chaos around the tracked
object. In order to overcome this situation, a context-based CF tracking framework (CA frame-
work) is introduced. In the learning stage, context blocks are added around the tracking target
by adding context information to the filter. The formula can be expressed as follows:
X
k
minkA0 w yk22 þ λ1 kwk22 þ λ2 kAi wk22 ð3Þ
w
i¼1
where Ai is the cyclic matrix corresponding to context patch. λ2 is a context patch regression
parameter. SAMF (CA) is an algorithm based on Equation 3:
yk22 þ λ1 kwk22
fp ðw; BÞ ¼ kBw ð4Þ
where
14
2 3 2 3
Affi0
pffiffiffiffi y
6 λ 2 A1 7 607
6 7
B¼6 .. 7and y¼6 7
4 ... 5: ð5Þ
4 . 5
pffiffiffiffiffi
λ2 Ak 0
Because fp ðw; BÞ is convex, it can be minimized by setting the gradient to zero, yielding:
w ¼ ðBT B þ λ1 IÞ1 BT
y ð5Þ
We linearly combine the new filter with the old one as below:
T
¼ θTnew þ ð1 θÞT ð6Þ
3 TRACKING COMPONENTS
15
Figure 1. Adaptive updating framework.
In order to adapt to the similarity between frames, an update rate related to image similarity
is constructed. We considering the relationship between each frame and the first frame of
target tracking image. If the similarity between the current frame and the first frame is high
and the change of tracking target is small, the correlation filter can be updated with a smaller
update rate. If the similarity between the current frame and the first frame is low, the change
of target may be large, and the correlation filter can be updated with a larger update rate. The
relationship between update rate and image frame similarity is proposed as follows:
1
θ¼ ð7Þ
1 þ ð5 þ mdHash Þ3
where θ is the update rate and mdHash is the mean of the difference hash value. According to
Equation 7, we can realize adaptive update rate.
16
Another random document with
no related content on Scribd:
circles. We must not, therefore, be surprised when a tragical tale of
this date brings before us the fact that Patrick Lindsay, described as
heir-male of the grand old House of Lindsay of the Byres, and who, a
few years afterwards, married a daughter of the sixteenth Earl of
Crawford, was now an upholsterer in the Parliament Close of
Edinburgh, and dean of guild for the city. Neither ought it to appear
as incredible that one of his apprentices was a youth named Cairns,
younger son of a gentleman of good estate residing at Cupar-Fife.
The tale was simply this—that, on the evening noted, between
eight and nine o’clock, Cairns was found in the shop expiring from
the effects of a violent blow on the head, apparently inflicted by a
hammer, while the box containing the guildry treasure was missing.
It was believed that some vile people who then haunted the city,
knowing of the box being kept in Lindsay’s shop, had formed a
design to possess themselves of it, and had effected their end at the
expense of murder, at the moment when the place was about to be
closed for the night. A number of vagrants were taken up on
suspicion, and the box was soon after found, empty.[679]
The engineer could not but wonder at Hill taking leave of the country
in this strain, ‘after he had been so exceedingly complaisant to it,
when here, as to compare its subterranean 1728.
riches with those of Mexico and Peru.’